Tiger Team: tt-copulas
Simulation study - A new variance estimator for empirical copulas in high dimensions
The Tiger Team scales an existing simulation study written in Python to the BWUniCluster 2.0. The simulation study aims to investigate the accuracy of an estimator for the variance of special empirical distribution functions, i.e. copulas, in high dimensions. The new estimator then can be used to construct statistical tests and improve various machine learning applications.
Members of the Tiger-Team:
Fabian Kächele, Institut für Operations Research, KIT; Hendrik Winkhardt, HPC Competence Center MLS&WISO